How stringent are

environmental policies?
POLICY PERSPECTIVES
 POLICY PERSPECTIVES
February 2016

How stringent are

environmental policies?

A review of OECD work on

indicators of environmental
policy stringency (EPS)

It is a common perception that environmental policies have become

increasingly stringent in response to environmental challenges, both global
and local. However, cross-country analysis of the economic effects of
environmental policies is held back by the lack of reliable, comparable
measures of stringency. This brochure presents new quantitative measures
of environmental policy stringency (EPS) developed by the OECD. The
EPS indicator is a composite index, derived through the aggregation of
information on selected environmental policy instruments, primarily related
to climate and air pollution. It covers most OECD countries over 1990-2012,
and has recently been extended to BRIICS (Brazil, Russia, India, Indonesia,
China and South Africa). While the EPS indicator is a simplification of the
multidimensional reality of environmental policies, it is a first tangible effort
to measure the stringency of environmental policies internationally over a
relatively long time horizon. It shows relatively high and significant correlations
with alternative proxies of EPS used in the literature, such as measures of
perceived stringency based on survey responses, outcomes and energy
prices etc. This brochure provides a description of the OECD approach, the
results and its properties as well as empirical applications and ongoing work
on this topic.

Contact information

Tomasz.Kozluk@oecd.org
Gregoire.Garsous@oecd.org
 POLICY PERSPECTIVES
February 2016

How stringent are

environmental policies?

A review of OECD work on

indicators of environmental
policy stringency (EPS)

It is a common perception that environmental policies have become

increasingly stringent in response to environmental challenges, both global
and local. However, cross-country analysis of the economic effects of
environmental policies is held back by the lack of reliable, comparable
measures of stringency. This brochure presents new quantitative measures
of environmental policy stringency (EPS) developed by the OECD. The
EPS indicator is a composite index, derived through the aggregation of
information on selected environmental policy instruments, primarily related
to climate and air pollution. It covers most OECD countries over 1990-2012,
and has recently been extended to BRIICS (Brazil, Russia, India, Indonesia,
China and South Africa). While the EPS indicator is a simplification of the
multidimensional reality of environmental policies, it is a first tangible effort
to measure the stringency of environmental policies internationally over a
relatively long time horizon. It shows relatively high and significant correlations
with alternative proxies of EPS used in the literature, such as measures of
perceived stringency based on survey responses, outcomes and energy
prices etc. This brochure provides a description of the OECD approach, the
results and its properties as well as empirical applications and ongoing work
on this topic.

Contact information

Tomasz.Kozluk@oecd.org
Gregoire.Garsous@oecd.org

What is environmental policy
stringency (EPS)?
Environmental policies contribute to wellbeing and the long-term
sustainability of growth. They strive to achieve environmental
objectives that markets fail to deliver. Such policies tend to make
pollution and, more generally environmental services, more costly
in order to change both producer and consumer behaviour. In
this respect, and for the purpose of this work, stringency can be
defined as the strength of the environmental policy signal – the
explicit or implicit cost of environmentally harmful behaviour, for
example pollution.
The stringency of environmental policies has been hypothesised
to have effects on economic activity, such as, for instance,
competitiveness and innovation. Still, actual empirical insights on
the effects of stringent policies have been weak and focused on
specific policies in specific contexts (Kozluk and Zipperer, 2014).
This has left the door open for interpretations that can often serve
the interest of the firms and sectors targeted by the policies, to the
detriment of economy-wide concerns – potentially both economic
and environmental.
To respond to this issue, the OECD has collected data on selected
environmental policies over countries and time in order to create a
proxy of environmental policy stringency and check its effects on
economic performance.
3 . © OECD how stringent are environmental policies?
Our approach:
The EPS Indicator
The OECD’s environmental policy stringency (EPS)
indicator aggregates information on selected
environmental policies to create a composite
measure of relative policy stringency across countries
and over time (Botta and Kozluk, 2014) . The indicator
focuses on upstream sectors, such as energy and
transport, of prime environmental importance and
similar relevance across countries. The policies
covered are environmentally-related taxes, renewable
energy and energy efficiency support (feed-in-tariffs,
renewable energy certificates, R&D expenditures),
performance standards (emission limit values for
coal fired power plants and sulphur content limits in
diesel fuels) and information on deposit and refund
schemes. Currently, the indicator focuses primarily
on air and climate policies.
The indicator is scored on a 0 to 6 scale, where 6
denotes most stringent policies. For each underlying
policy instrument, stringency is defined as a higher
implicit or explicit price placed on the relevant
environmental damage produced by firms or
Data sources and relevant links
EPS indicator website: http://oe.cd/eps
Environmental Policy Stringency data:
http://stats.oecd.org/Index.aspx?DataSetCode=EPS
POLICY PERSPECTIVES
consumers. Policies can alter the costs of pollution
in different ways. “Stick” type policies, that
directly raise the costs of polluting behaviour, are
considered more stringent the higher the taxes or
the stricter the standards. “Carrot” type policies,
which reward “environmentally-friendlier”
activity (thereby raising the relative costs of
pollution), are assumed more stringent the
higher the support such as feed-in-tariffs or R&D
subsidies. Page 5 presents the structure of the
EPS indicator including more technical details on
the policy instruments and aggregation. Overall,
the EPS measure should be treated and used
as a proxy for aggregate environmental policy
stringency. The following sections show how
countries score on the EPS indicator and discuss
the properties, advantages, disadvantages and
challenges of such an approach.
© OECD how stringent are environmental policies? . 4

What is environmental policy
stringency (EPS)?
Environmental policies contribute to wellbeing and the long-term
sustainability of growth. They strive to achieve environmental
objectives that markets fail to deliver. Such policies tend to make
pollution and, more generally environmental services, more costly
in order to change both producer and consumer behaviour. In
this respect, and for the purpose of this work, stringency can be
defined as the strength of the environmental policy signal – the
explicit or implicit cost of environmentally harmful behaviour, for
example pollution.
The stringency of environmental policies has been hypothesised
to have effects on economic activity, such as, for instance,
competitiveness and innovation. Still, actual empirical insights on
the effects of stringent policies have been weak and focused on
specific policies in specific contexts (Kozluk and Zipperer, 2014).
This has left the door open for interpretations that can often serve
the interest of the firms and sectors targeted by the policies, to the
detriment of economy-wide concerns – potentially both economic
and environmental.
To respond to this issue, the OECD has collected data on selected
environmental policies over countries and time in order to create a
proxy of environmental policy stringency and check its effects on
economic performance.
3 . © OECD how stringent are environmental policies?
Our approach:
The EPS Indicator
The OECD’s environmental policy stringency (EPS)
indicator aggregates information on selected
environmental policies to create a composite
measure of relative policy stringency across countries
and over time (Botta and Kozluk, 2014) . The indicator
focuses on upstream sectors, such as energy and
transport, of prime environmental importance and
similar relevance across countries. The policies
covered are environmentally-related taxes, renewable
energy and energy efficiency support (feed-in-tariffs,
renewable energy certificates, R&D expenditures),
performance standards (emission limit values for
coal fired power plants and sulphur content limits in
diesel fuels) and information on deposit and refund
schemes. Currently, the indicator focuses primarily
on air and climate policies.
The indicator is scored on a 0 to 6 scale, where 6
denotes most stringent policies. For each underlying
policy instrument, stringency is defined as a higher
implicit or explicit price placed on the relevant
environmental damage produced by firms or
Data sources and relevant links
EPS indicator website: http://oe.cd/eps
Environmental Policy Stringency data:
http://stats.oecd.org/Index.aspx?DataSetCode=EPS
POLICY PERSPECTIVES
consumers. Policies can alter the costs of pollution
in different ways. “Stick” type policies, that
directly raise the costs of polluting behaviour, are
considered more stringent the higher the taxes or
the stricter the standards. “Carrot” type policies,
which reward “environmentally-friendlier”
activity (thereby raising the relative costs of
pollution), are assumed more stringent the
higher the support such as feed-in-tariffs or R&D
subsidies. Page 5 presents the structure of the
EPS indicator including more technical details on
the policy instruments and aggregation. Overall,
the EPS measure should be treated and used
as a proxy for aggregate environmental policy
stringency. The following sections show how
countries score on the EPS indicator and discuss
the properties, advantages, disadvantages and
challenges of such an approach.
© OECD how stringent are environmental policies? . 4


POLICY PERSPECTIVES





THE STRUCTURE OF THE EPS INDICATOR Environmental policy

The overall structure of the EPS index is presented in Figure 1. For each individual instrument, the quantitative
stringency in OECD
or qualitative information is normalised where relevant (e.g. taxes on pollutants from the electricity sector are


rescaled using electricity prices). They are then scored on a 0-6 scale increasing in stringency (0 is assigned when

the instrument is not present in a country). The thresholds for each bin are determined based on the stringency
countries and BRIICS

of the given measures across countries and time. The country scores are then aggregated by instrument type

(taxes, trading schemes, emission standards, etc.), instrument category (market-based and non-market based) and


further on using equal weights at each stage. In practice, the country rankings are relatively robust to alternative

weighting and aggregation. The indicator covers the years 1990-2012 for 24 OECD countries. For federal countries,

where some of the key instruments are applied at the sub-national levels, the national indicator is a weighted

average of regional policies, where weights are the share of each region in electricity consumption/production.
Figure 2 displays the EPS indicator scores of OECD over the past two decades.
countries for the period 1990-1995 and the year • Environmental policies remain more

In 2015, the EPS indicator was extended to cover BRIICS (Brazil, Russia, India, Indonesia, China and South Africa). 2012. Figures 3 and 4 show the performance stringent in OECD countries than in BRIICS.

This required two minor modifications due to data constraints (Kozluk and Timiliotis (2016)). First, the subindicator
of countries on the extended EPS indicator
on deposit and refund schemes was excluded. All other subindicators were reweighted accordingly. Second, which includes BRIICS. Several patterns can be • Policies, as measured by the EPS indicator,

public expenditures on R&D in renewable energy for the BRIICS were interpolated at low levels (score of 1). Over
distinguished: are most stringent in Nordic countries, the

the entire sample of OECD countries, the extended EPS indicator is highly correlated to the original one, with a Netherlands, Finland and Germany. Among

correlation of 0.92 (significant at 1%). • Environmental policy stringency has been OECD countries, they are least stringent in

increasing in all OECD countries and BRIICS Greece, Portugal, Ireland and Hungary. Most
Figure 1. Aggregation structure of the composite index of EPS of the other countries are close to the
OECD average.

Composite indicator of Figure 2. Environmental policy stringency in OECD countries

environmental policy stringency

0.5 0.5 Indicator value - 1990/1995 Indicator value - 2012 OECD average - 2012
5
More stringent
Market-based policies Non-market based policies
4
0.25 0.25 0.25 0.25 0.5 0.5

3
Trading R&D
Taxes FITs DRS Standards
Schemes Subsidies

• CO2 Emission Limit 2

• Renewable Government
Values:
• CO2 Energy R&D
• Deposit • NOx
• NOx Certificates • Solar expenditure
& Refund • SOx
• SOx • Energy • Wind on
Scheme • PMx 1
• Diesel Efficiency Renewable
• Sulphur content
Certificates Energy
limit (Diesel)

Note: FITs encompasses feed-in tariffs and feed-in premiums for renewable energy (wind and solar).
Source: Botta and Kozluk, (2014); OECD/EEA Environmental Policy Instruments www.oecd.org/env/policies/da-
tabase; OECD-EPAU dataset on Renewable Energy Policies www.oecd.org/env/consumption-innovation/finance.
htm.
Source: Botta and Kozluk (2014). Figure 3 displays the EPS indicator scores of OECD countries for the period
1990-1995 and the year 2012.

5 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 6


POLICY PERSPECTIVES





THE STRUCTURE OF THE EPS INDICATOR Environmental policy

The overall structure of the EPS index is presented in Figure 1. For each individual instrument, the quantitative
stringency in OECD
or qualitative information is normalised where relevant (e.g. taxes on pollutants from the electricity sector are


rescaled using electricity prices). They are then scored on a 0-6 scale increasing in stringency (0 is assigned when

the instrument is not present in a country). The thresholds for each bin are determined based on the stringency
countries and BRIICS

of the given measures across countries and time. The country scores are then aggregated by instrument type

(taxes, trading schemes, emission standards, etc.), instrument category (market-based and non-market based) and


further on using equal weights at each stage. In practice, the country rankings are relatively robust to alternative

weighting and aggregation. The indicator covers the years 1990-2012 for 24 OECD countries. For federal countries,

where some of the key instruments are applied at the sub-national levels, the national indicator is a weighted

average of regional policies, where weights are the share of each region in electricity consumption/production.
Figure 2 displays the EPS indicator scores of OECD over the past two decades.
countries for the period 1990-1995 and the year • Environmental policies remain more

In 2015, the EPS indicator was extended to cover BRIICS (Brazil, Russia, India, Indonesia, China and South Africa). 2012. Figures 3 and 4 show the performance stringent in OECD countries than in BRIICS.

This required two minor modifications due to data constraints (Kozluk and Timiliotis (2016)). First, the subindicator
of countries on the extended EPS indicator
on deposit and refund schemes was excluded. All other subindicators were reweighted accordingly. Second, which includes BRIICS. Several patterns can be • Policies, as measured by the EPS indicator,

public expenditures on R&D in renewable energy for the BRIICS were interpolated at low levels (score of 1). Over
distinguished: are most stringent in Nordic countries, the

the entire sample of OECD countries, the extended EPS indicator is highly correlated to the original one, with a Netherlands, Finland and Germany. Among

correlation of 0.92 (significant at 1%). • Environmental policy stringency has been OECD countries, they are least stringent in

increasing in all OECD countries and BRIICS Greece, Portugal, Ireland and Hungary. Most
Figure 1. Aggregation structure of the composite index of EPS of the other countries are close to the
OECD average.

Composite indicator of Figure 2. Environmental policy stringency in OECD countries

environmental policy stringency

0.5 0.5 Indicator value - 1990/1995 Indicator value - 2012 OECD average - 2012
5
More stringent
Market-based policies Non-market based policies
4
0.25 0.25 0.25 0.25 0.5 0.5

3
Trading R&D
Taxes FITs DRS Standards
Schemes Subsidies

• CO2 Emission Limit 2

• Renewable Government
Values:
• CO2 Energy R&D
• Deposit • NOx
• NOx Certificates • Solar expenditure
& Refund • SOx
• SOx • Energy • Wind on
Scheme • PMx 1
• Diesel Efficiency Renewable
• Sulphur content
Certificates Energy
limit (Diesel)

Note: FITs encompasses feed-in tariffs and feed-in premiums for renewable energy (wind and solar).
Source: Botta and Kozluk, (2014); OECD/EEA Environmental Policy Instruments www.oecd.org/env/policies/da-
tabase; OECD-EPAU dataset on Renewable Energy Policies www.oecd.org/env/consumption-innovation/finance.
htm.
Source: Botta and Kozluk (2014). Figure 3 displays the EPS indicator scores of OECD countries for the period
1990-1995 and the year 2012.

5 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 6
 POLICY PERSPECTIVES
Figure 3. Environmental policy stringency in OECD and BRIICS The EPS indicator has also been compared with various proxies of
environmental performance. Table 1 highlights the correlations with
CO2 intensity of the economy and electricity generation as well as
1990-95 2012 the Yale Environmental Performance Index (EPI).
4

3.5 It also shows a significant positive correlation with GDP per capita,
confirming that richer countries tend to have more stringent
3 policies.

2.5
More stringent

1.5

1 Table 1. Correlations with emission intensity and environmental performance

0.5

0 CO2/GDP CO2/KWh EPI

BRIICS Lowest OECD Highest OECD

EPS -0.41 -0.36 0.26

Note: BRIICS countries are Brazil, Russia, India, Indonesia, China and South-Africa. Lowest OECD countries are
Greece, Ireland and Portugal. Highest OECD countries are Finland, Netherlands and Denmark.

Note: Spearman correlations calculated with EPS indicator for OECD countries and BRIICS for the period 1990-
2012. All correlations are significant at 1% level.

Figure 4. Environmental policy stringency in BRIICS

1990-95 2012

1.4

1.2

1
More stringent

0.8

0.6

0.4

0.2

0
Russia Brazil South Africa Indonesia India China

7 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 8
 POLICY PERSPECTIVES
Figure 3. Environmental policy stringency in OECD and BRIICS The EPS indicator has also been compared with various proxies of
environmental performance. Table 1 highlights the correlations with
CO2 intensity of the economy and electricity generation as well as
1990-95 2012 the Yale Environmental Performance Index (EPI).
4

3.5 It also shows a significant positive correlation with GDP per capita,
confirming that richer countries tend to have more stringent
3 policies.

2.5
More stringent

1.5

1 Table 1. Correlations with emission intensity and environmental performance

0.5

0 CO2/GDP CO2/KWh EPI

BRIICS Lowest OECD Highest OECD

EPS -0.41 -0.36 0.26

Note: BRIICS countries are Brazil, Russia, India, Indonesia, China and South-Africa. Lowest OECD countries are
Greece, Ireland and Portugal. Highest OECD countries are Finland, Netherlands and Denmark.

Note: Spearman correlations calculated with EPS indicator for OECD countries and BRIICS for the period 1990-
2012. All correlations are significant at 1% level.

Figure 4. Environmental policy stringency in BRIICS

1990-95 2012

1.4

1.2

1
More stringent

0.8

0.6

0.4

0.2

0
Russia Brazil South Africa Indonesia India China

7 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 8
 POLICY PERSPECTIVES
Advantages and disadvantages

The main advantage of the EPS index is that it The main issue with these types of instruments is
is a simple proxy, based on actual policies. It the high level of site specificity and the difficulty in
concentrates on climate and air policies in key assessing and comparing stringency. For instance,
upstream sectors – which are important and land use regulation is likely to be implemented
relatively comparable polluters across economies. at lower levels of governments while voluntary
approaches are often negotiated between single
The underlying assumption is that a stringent facilities and local authorities. This heterogeneity,
approach to environmental issues across a number Figure 5. Environmental policy stringency in OECD and BRIICS
together with the possibility of regulatory capture,
of key pollutants and industries is representative of makes the comparison of their stringency across
a more general preoccupation with environmental time and countries even more problematic. The
issues across all areas. omission of certain types of instruments weakens
Environmental Implementation Perceptions of Changes in Changes in
the generality of the composite indicator since policy Environmental agents' environmental
The policy instruments covered are comparable instruments Enforcement policies behaviour outcomes
these instruments are common in some countries, 3 4
in a relatively straightforward fashion in terms 1 2
like Japan, where voluntary approaches represent a
of stringency. The resulting indicator has broader E.g. Event based E.g. Shadow prices, E.g. Environmental
large share of the tools. approach, composite
E.g. Survey of
survey of firms performance data
cross-country and time coverage than other executives
abatement costs
indicators
available direct policy measures. It can also be In light of these limitations, it is useful to compare
easily updated and expanded once relevant the EPS indicator with other attempts to measure Environmental laws are designed … to alter firms' and achieve
and implemented…. (and consumers')
data become available. Moreover, in practice, the environmental policies across countries and time "cleaner"outcome
behaviour
indicator is fairly stable with respect to changes that are used in the empirical literature. Such
and weighting and aggregation. However, such a approaches differ in many dimensions, but it is
simple proxy has numerous limitations, due to possible to classify them according to “where” they
the many challenges in measuring environmental attempt to measure environmental stringency (as Table 2. Correlations of extended EPS indicator with other proxies of environmental policy
policy stringency (see p.12). represented by the bubbles in Figure 5). stringency

First, the focus on limited environmental policy Spearman rank correlations over maximum available sample
areas overlooks other important areas such as
water, biodiversity, natural resources or waste. Correlation with
EPS measure Sample characteristics
EPS
Extending the policies covered will be the subject of
Original EPS 0.92*** OECD countries only, 1990-2012
future work.
CLIMI (EBRD) 0.45** OECD countries and BRIICS, 2008

Second, even in the areas of air and climate, World Economic Forum’s Executive OECD countries and BRIICS, 2002-
0.50***
Opinion Survey 2012
some potentially relevant policy instruments
OECD countries and BRIICS, 1990-
have been ignored. For example, tax incentives Energy Prices (Sato et al. 2015) 0.50***
2012
for “environmentally-friendly” investment, OECD countries and BRIICS, 1990-
Environmental Patents (share) 0.28***
land use regulations or labelling obligations, 2012

voluntary approaches (VAs) and other “soft” policy OECD countries and BRIICS excluding
Landfilled waste (share, Sauvage 2014) -0.59***
Indonesia, 1995-2012
instruments are not covered.
Note: ***, **, * denote significance at 1%, 5% and 10% respectively.

9 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 10
 POLICY PERSPECTIVES
Advantages and disadvantages

The main advantage of the EPS index is that it The main issue with these types of instruments is
is a simple proxy, based on actual policies. It the high level of site specificity and the difficulty in
concentrates on climate and air policies in key assessing and comparing stringency. For instance,
upstream sectors – which are important and land use regulation is likely to be implemented
relatively comparable polluters across economies. at lower levels of governments while voluntary
approaches are often negotiated between single
The underlying assumption is that a stringent facilities and local authorities. This heterogeneity,
approach to environmental issues across a number Figure 5. Environmental policy stringency in OECD and BRIICS
together with the possibility of regulatory capture,
of key pollutants and industries is representative of makes the comparison of their stringency across
a more general preoccupation with environmental time and countries even more problematic. The
issues across all areas. omission of certain types of instruments weakens
Environmental Implementation Perceptions of Changes in Changes in
the generality of the composite indicator since policy Environmental agents' environmental
The policy instruments covered are comparable instruments Enforcement policies behaviour outcomes
these instruments are common in some countries, 3 4
in a relatively straightforward fashion in terms 1 2
like Japan, where voluntary approaches represent a
of stringency. The resulting indicator has broader E.g. Event based E.g. Shadow prices, E.g. Environmental
large share of the tools. approach, composite
E.g. Survey of
survey of firms performance data
cross-country and time coverage than other executives
abatement costs
indicators
available direct policy measures. It can also be In light of these limitations, it is useful to compare
easily updated and expanded once relevant the EPS indicator with other attempts to measure Environmental laws are designed … to alter firms' and achieve
and implemented…. (and consumers')
data become available. Moreover, in practice, the environmental policies across countries and time "cleaner"outcome
behaviour
indicator is fairly stable with respect to changes that are used in the empirical literature. Such
and weighting and aggregation. However, such a approaches differ in many dimensions, but it is
simple proxy has numerous limitations, due to possible to classify them according to “where” they
the many challenges in measuring environmental attempt to measure environmental stringency (as Table 2. Correlations of extended EPS indicator with other proxies of environmental policy
policy stringency (see p.12). represented by the bubbles in Figure 5). stringency

First, the focus on limited environmental policy Spearman rank correlations over maximum available sample
areas overlooks other important areas such as
water, biodiversity, natural resources or waste. Correlation with
EPS measure Sample characteristics
EPS
Extending the policies covered will be the subject of
Original EPS 0.92*** OECD countries only, 1990-2012
future work.
CLIMI (EBRD) 0.45** OECD countries and BRIICS, 2008

Second, even in the areas of air and climate, World Economic Forum’s Executive OECD countries and BRIICS, 2002-
0.50***
Opinion Survey 2012
some potentially relevant policy instruments
OECD countries and BRIICS, 1990-
have been ignored. For example, tax incentives Energy Prices (Sato et al. 2015) 0.50***
2012
for “environmentally-friendly” investment, OECD countries and BRIICS, 1990-
Environmental Patents (share) 0.28***
land use regulations or labelling obligations, 2012

voluntary approaches (VAs) and other “soft” policy OECD countries and BRIICS excluding
Landfilled waste (share, Sauvage 2014) -0.59***
Indonesia, 1995-2012
instruments are not covered.
Note: ***, **, * denote significance at 1%, 5% and 10% respectively.

9 . © OECD how stringent are environmental policies? © OECD how stringent are environmental policies? . 10

A first set of measures attempts to capture actual
taxes, laws and regulations directly. They are the
closest to actual policy instruments and include
indicators of existence of single policies, their levels
(e.g. tax rates) or changes as well as composite
measures that aggregate selected information on
individual instruments. The main assumption is
that it is possible to directly observe environmental
regulations, or at least representative elements
of them, and that it is possible to represent a
country’s stance on environmental regulation by
summarising selected measures of the enforced
laws (Figure 5, bubble 1). Measures attempting
to capture perceptions of the stringency of
environmental policies are another set, based on
dedicated survey questionnaires (bubble 2).
A third set of measures focuses on the first-order
consequences of environmental regulations –such
as firms’ costs, actions and production choices
(bubble 3). They include estimates of shadow prices
for environmental inputs or environmentally-
related expenditures (e.g. on pollution control and
abatement), often self-reported.
11 . © OECD how stringent are environmental policies?
Finally, environmental performance measures
look at the second-level consequences
of instruments, that is, the variation in
environmental performance of firms, sectors or
countries, in order to proxy the stringency of the
policy itself (bubble 4).
The EPS indicator aims to capture the stringency
of actual policies, so the most relevant
comparisons are those from the first and second
of the above categories. These also happen to
be the measures most widely available across
countries and time, even if samples differ from
those of the EPS indicator. They include the
EBRD’s Climate Laws, Institutions and Measures
Index and the World Economic Forum’s Executive
Opinion Survey responses (managers’ perceptions
on domestic policy stringency). In addition,
checks with energy prices, also used as proxies
of environmental policy stringency (Sato et al.,
2015) and proxies for environmentally-related
innovation (“green” patent shares) are provided.
As shown in Table 2, correlations are significant,
suggesting that each of these measures captures
part of the same phenomenon. Each single
measure is far from ideal; hence in empirical
work checks with alternative measures of
environmental policy stringency can be valuable.
POLICY
POLICY PERSPECTIVES
PERSPECTIVES
Challenges in measuring the stringency of
environmental policies
Measuring environmental policy stringency is a cumbersome task, not least because of the lack of
appropriate data. The main challenges are due to the heterogeneous nature of environmental issues
and the various instruments that address them (Brunel and Levinson, 2013; Kozluk and Zipperer,
2014):
Multi-dimensionality is due to the various dimensions of environmental regulations (e.g. across
different media, such as air, water, soil), the multitude of available policy instruments (e.g. pricing,
command and control instruments, voluntary approaches) and their design and implementation
features. Legislators may regulate emissions of a single pollutant through different instruments (e.g.
both a tax and a performance standard on NOx emissions), but often also discriminate the application
of regulations according to the sectors, technologies, vintages or sizes of plants or location of activity
(e.g. urban area). Environmental regulation can also be implemented at various levels of government.
Sampling arises as the sample of industries subject to policies may be driven by the policies
themselves. For example, more polluting industries may represent a lower share of GDP in a country
subject to stringent policies precisely because the policies have shifted the industrial structure away
from them. Moreover, in sectors indirectly affected by stringent environmental policies (e.g. the
service sector may be affected primarily through high electricity prices), these effects are not likely to
be correctly assessed as resulting from environmental policies.
Identification is the difficulty in correctly assessing the degree to which the expected consequences
of stricter regulations (e.g. increased abatement expenditures or observed lower pollution intensity)
can be actually attributed to environmental policy stringency. Observed environmental outcomes can
be due to other regulatory instruments (for instance, affecting labour and capital) and country-specific
characteristics: geography, market imperfections, skills, level of development, access to technology
or trade openness and outsourcing. All these features tend to interact with each other, making it
difficult to link measures of relative environmental performance (or abatement expenditures) to actual
environmental policies.
Implementation and enforcement are linked to broader characteristics of environmental policies,
including their flexibility, depth, predictability and competition-friendliness. Further effects may come
from the provisions accompanying the environmental policies (e.g. to smooth the transition) and the
discrepancy between the legal and the de facto stringency: the transposition of laws into government
actions, levels of fines and lax pursuit of violations. Enforcement may be of particular importance in
countries with lower quality of institutions or large unofficial economies – for example in developing
countries (Scrieciu, 2015).
© OECD how stringent are environmental policies? . 12

A first set of measures attempts to capture actual
taxes, laws and regulations directly. They are the
closest to actual policy instruments and include
indicators of existence of single policies, their levels
(e.g. tax rates) or changes as well as composite
measures that aggregate selected information on
individual instruments. The main assumption is
that it is possible to directly observe environmental
regulations, or at least representative elements
of them, and that it is possible to represent a
country’s stance on environmental regulation by
summarising selected measures of the enforced
laws (Figure 5, bubble 1). Measures attempting
to capture perceptions of the stringency of
environmental policies are another set, based on
dedicated survey questionnaires (bubble 2).
A third set of measures focuses on the first-order
consequences of environmental regulations –such
as firms’ costs, actions and production choices
(bubble 3). They include estimates of shadow prices
for environmental inputs or environmentally-
related expenditures (e.g. on pollution control and
abatement), often self-reported.
11 . © OECD how stringent are environmental policies?
Finally, environmental performance measures
look at the second-level consequences
of instruments, that is, the variation in
environmental performance of firms, sectors or
countries, in order to proxy the stringency of the
policy itself (bubble 4).
The EPS indicator aims to capture the stringency
of actual policies, so the most relevant
comparisons are those from the first and second
of the above categories. These also happen to
be the measures most widely available across
countries and time, even if samples differ from
those of the EPS indicator. They include the
EBRD’s Climate Laws, Institutions and Measures
Index and the World Economic Forum’s Executive
Opinion Survey responses (managers’ perceptions
on domestic policy stringency). In addition,
checks with energy prices, also used as proxies
of environmental policy stringency (Sato et al.,
2015) and proxies for environmentally-related
innovation (“green” patent shares) are provided.
As shown in Table 2, correlations are significant,
suggesting that each of these measures captures
part of the same phenomenon. Each single
measure is far from ideal; hence in empirical
work checks with alternative measures of
environmental policy stringency can be valuable.
POLICY
POLICY PERSPECTIVES
PERSPECTIVES
Challenges in measuring the stringency of
environmental policies
Measuring environmental policy stringency is a cumbersome task, not least because of the lack of
appropriate data. The main challenges are due to the heterogeneous nature of environmental issues
and the various instruments that address them (Brunel and Levinson, 2013; Kozluk and Zipperer,
2014):
Multi-dimensionality is due to the various dimensions of environmental regulations (e.g. across
different media, such as air, water, soil), the multitude of available policy instruments (e.g. pricing,
command and control instruments, voluntary approaches) and their design and implementation
features. Legislators may regulate emissions of a single pollutant through different instruments (e.g.
both a tax and a performance standard on NOx emissions), but often also discriminate the application
of regulations according to the sectors, technologies, vintages or sizes of plants or location of activity
(e.g. urban area). Environmental regulation can also be implemented at various levels of government.
Sampling arises as the sample of industries subject to policies may be driven by the policies
themselves. For example, more polluting industries may represent a lower share of GDP in a country
subject to stringent policies precisely because the policies have shifted the industrial structure away
from them. Moreover, in sectors indirectly affected by stringent environmental policies (e.g. the
service sector may be affected primarily through high electricity prices), these effects are not likely to
be correctly assessed as resulting from environmental policies.
Identification is the difficulty in correctly assessing the degree to which the expected consequences
of stricter regulations (e.g. increased abatement expenditures or observed lower pollution intensity)
can be actually attributed to environmental policy stringency. Observed environmental outcomes can
be due to other regulatory instruments (for instance, affecting labour and capital) and country-specific
characteristics: geography, market imperfections, skills, level of development, access to technology
or trade openness and outsourcing. All these features tend to interact with each other, making it
difficult to link measures of relative environmental performance (or abatement expenditures) to actual
environmental policies.
Implementation and enforcement are linked to broader characteristics of environmental policies,
including their flexibility, depth, predictability and competition-friendliness. Further effects may come
from the provisions accompanying the environmental policies (e.g. to smooth the transition) and the
discrepancy between the legal and the de facto stringency: the transposition of laws into government
actions, levels of fines and lax pursuit of violations. Enforcement may be of particular importance in
countries with lower quality of institutions or large unofficial economies – for example in developing
countries (Scrieciu, 2015).
© OECD how stringent are environmental policies? . 12
 POLICY
POLICY PERSPECTIVES
Improvements and extensions

The EPS index is work in progress. Despite a
number of favourable properties, the coverage
remains limited - focusing on air and climate
pollutants largely in upstream activities such as
electricity production and transport.
While this was a necessary assumption when the
indicator was conceived, a natural next step is to
look at the possibility of including other major
environmental policies. Scoping work to include
policies for regulating water pollution is currently
PERSPECTIVES
ongoing. In parallel, work is being conducted on
extending the coverage to other OECD countries
not yet covered and further improving the coverage
of selected instruments. This work is expected to
yield results by late 2016. In light of the recent
extensions, Table 3 shows the progress of the data
collection process for BRIICS countries.

Empirical applications no evidence that stringent environmental

policies harm aggregate trade and overall
Table 3. Mapping of policies covered for BRIICS (so far)

The OECD’s EPS index has already been used in
empirical work in order to measure the effect of
environmental policy stringency on economic
outcomes:
• Albrizio et al. (2014) show that the
tightening of environmental policies
observed in OECD countries had little
effect on aggregate productivity growth.
Nevertheless, they find that increasing EPS
has led to differentiated effects within
the economy – the most technologically
advanced industries and firms have seen a
small, temporary increase in productivity
growth whereas the least productive firms
have seen their productivity fall. The
possible channels for this are firm dynamics
(entry/exit and growth), outsourcing and
innovation.
• Kozluk and Timiliotis (2016) investigate
the Pollution Haven Hypothesis - whether
countries with more stringent environmental
policies lose out in terms of competitiveness
and exports. Using the extended EPS
indicator which covers BRIICS, they compare
effects on gross exports with those on
domestic value added embedded in exports
– a Global Value Chain approach. They find
country competitiveness. However,
environmental policies are found to have Tax Trading schemes FIT ELV R&D
a significant effect on trade specialisation. CO2 NOX SOX Die- Green CO2 White SO2 Wind Solar SOX NOX PM Sulphur Subsidies
sel
More stringent environmental policies are
Brazil
associated with a competitiveness loss in
Russia
the most polluting sectors and a reduction
in their exports. At the same time they are India
linked to a comparative advantage in less Indonesia
polluting industries and a boost in exports. China
Overall, the effects of environmental policies
South Africa
on trade patterns over the past two decades
have been found small with respect to The information for this variable/country pair is completely collected and up to date.
effects of other developments, such as trade The information for this variable/country pair is partially collected with missing values.
liberalisation. The information for this variable/country pair is not available or does not exist.
• In a similar vein, Sauvage (2014) finds a
significant positive relationship between
a country’s regulatory stringency and its
specialisation in exports of “environmental”
products.
• Other researchers have used the EPS index PHOTO CREDITS
to test the effects on innovation (De Santis
© SOMMAI/Shutterstock.com © Gyuszko-Photo/Shutterstock.com
and Lasinio, 2015); to measure how policy © SusaZoom/Shutterstock.com © Alex Mit/Shutterstock.com
stringency affects CO2 emissions (Probst and © Michal Staniewski/Shutterstock.com © Robert Neumann/Shutterstock.com
Sauter, 2015); and to estimate the learning © apiguide/Shutterstock.com © studio23/Shutterstock.com
curve of renewable energy technologies © Peter Gudella/Shutterstock.com © iStockphoto.com/gashgeron
© manfredxy/Shutterstock.com ©Julia Ivantsova/Shutterstock.com
(Witajewski-Baltvilksa et al., 2015).

13 . © OECD how stringent are environmental policies?

© OECD how stringent are environmental policies? . 14
 POLICY
POLICY PERSPECTIVES
Improvements and extensions

The EPS index is work in progress. Despite a
number of favourable properties, the coverage
remains limited - focusing on air and climate
pollutants largely in upstream activities such as
electricity production and transport.
While this was a necessary assumption when the
indicator was conceived, a natural next step is to
look at the possibility of including other major
environmental policies. Scoping work to include
policies for regulating water pollution is currently
PERSPECTIVES
ongoing. In parallel, work is being conducted on
extending the coverage to other OECD countries
not yet covered and further improving the coverage
of selected instruments. This work is expected to
yield results by late 2016. In light of the recent
extensions, Table 3 shows the progress of the data
collection process for BRIICS countries.

Empirical applications no evidence that stringent environmental

policies harm aggregate trade and overall
Table 3. Mapping of policies covered for BRIICS (so far)

The OECD’s EPS index has already been used in
empirical work in order to measure the effect of
environmental policy stringency on economic
outcomes:
• Albrizio et al. (2014) show that the
tightening of environmental policies
observed in OECD countries had little
effect on aggregate productivity growth.
Nevertheless, they find that increasing EPS
has led to differentiated effects within
the economy – the most technologically
advanced industries and firms have seen a
small, temporary increase in productivity
growth whereas the least productive firms
have seen their productivity fall. The
possible channels for this are firm dynamics
(entry/exit and growth), outsourcing and
innovation.
• Kozluk and Timiliotis (2016) investigate
the Pollution Haven Hypothesis - whether
countries with more stringent environmental
policies lose out in terms of competitiveness
and exports. Using the extended EPS
indicator which covers BRIICS, they compare
effects on gross exports with those on
domestic value added embedded in exports
– a Global Value Chain approach. They find
country competitiveness. However,
environmental policies are found to have Tax Trading schemes FIT ELV R&D
a significant effect on trade specialisation. CO2 NOX SOX Die- Green CO2 White SO2 Wind Solar SOX NOX PM Sulphur Subsidies
sel
More stringent environmental policies are
Brazil
associated with a competitiveness loss in
Russia
the most polluting sectors and a reduction
in their exports. At the same time they are India
linked to a comparative advantage in less Indonesia
polluting industries and a boost in exports. China
Overall, the effects of environmental policies
South Africa
on trade patterns over the past two decades
have been found small with respect to The information for this variable/country pair is completely collected and up to date.
effects of other developments, such as trade The information for this variable/country pair is partially collected with missing values.
liberalisation. The information for this variable/country pair is not available or does not exist.
• In a similar vein, Sauvage (2014) finds a
significant positive relationship between
a country’s regulatory stringency and its
specialisation in exports of “environmental”
products.
• Other researchers have used the EPS index PHOTO CREDITS
to test the effects on innovation (De Santis
© SOMMAI/Shutterstock.com © Gyuszko-Photo/Shutterstock.com
and Lasinio, 2015); to measure how policy © SusaZoom/Shutterstock.com © Alex Mit/Shutterstock.com
stringency affects CO2 emissions (Probst and © Michal Staniewski/Shutterstock.com © Robert Neumann/Shutterstock.com
Sauter, 2015); and to estimate the learning © apiguide/Shutterstock.com © studio23/Shutterstock.com
curve of renewable energy technologies © Peter Gudella/Shutterstock.com © iStockphoto.com/gashgeron
© manfredxy/Shutterstock.com ©Julia Ivantsova/Shutterstock.com
(Witajewski-Baltvilksa et al., 2015).

13 . © OECD how stringent are environmental policies?

© OECD how stringent are environmental policies? . 14
How stringent are
environmental policies?

REFERENCES
How stringent are
Albrizio, S., E. Botta, T.Kozluk and V. Zipperer (2014): “Do Environmental Policies Matter for
Productivity Growth? Insight from New Cross-Country Measures of Environmental Policies,”
OECD Economics Department Working Papers, No. 1176, OECD Publishing.
environmental policies?
Botta, E. and T. Kozluk (2014): “Measuring environmental policy stringency in OECD countries:
A composite index approach”, OECD Economics Department Working Papers, No. 1177,
OECD Publishing.

Brunel, C. and A. Levinson, (2013): “Measuring Environmental Regulatory Stringency”, OECD

POLICY PERSPECTIVES
Trade and Environment Working Papers 2013/5, OECD Publishing.

De Santis, R. and C. Lasinio (2015): “Environmental Policies, Innovation and Productivity in the
EU,” LEQS Paper No. 100/2015.

Kozluk, T. and V. Zipperer (2014), “Environmental Policies and Productivity Growth: A Critical
Review of Empirical Findings”, OECD Journal: Economic Studies, Vol. 1, 2014.

Kozluk, T. and C. Timiliotis (2016), “Do environmental policies affect global value chains? A
new perspective on the pollution haven hypothesis,” OECD Economics Department Working
Papers, No. 1282.

Probst, M. and C. Sauter (2015): “CO2 Emissions and Greenhouse Gas Policy Stringency: An
Empirical Assessment,” IRENE Working papers 15-03.

Sauvage, J. (2014): “The Stringency of Environmental Regulations and Trade in Environmental

Goods,” OECD Trade and Environment Working Papers, No. 2014/03, OECD Publishing.

Sato, M., G. Singer, D. Dussaux S. and Lovo (2015): “International and sectoral variation in
energy prices 1995-2011: how does it relate to emissions policy stringency?”, GRI Working
Papers 187, Grantham Research Institute on Climate Change and the Environment.

Scrieciu, S. (2015): “Measuring Environmental Action and Economic Performance in Developing

Countries,” GGKP Working Paper 01 | 2015.

Witajewski-Baltvilksa, J., E. Verdolinib and M. Tavonic (2015): “Bending the learning curve,”
Energy Economics, Vol. 52, Supplement 1, pp. S86–S99.

http://oe.cd/eps

Contact

Tomasz.Kozluk@oecd.org
Gregoire.Garsous@oecd.org

February 2016